The Agent Skills Directory

[EXTERNAL_DOWNLOADS]: The skill references the official Juniper repository (github.com/Juniper/junos-mcp-server) for its underlying MCP server implementation. As Juniper is a well-known networking vendor, this is a trusted source.
[COMMAND_EXECUTION]: Tools such as execute_junos_command and load_and_commit_config permit high-privilege operations, including CLI execution and configuration changes on remote network hardware. These capabilities are consistent with the skill's primary purpose of network automation.
[PROMPT_INJECTION]: The skill presents an indirect prompt injection surface as it processes arbitrary strings for CLI commands and Jinja2 templates. This risk is effectively mitigated by built-in safety features:
Ingestion points: Tools accept command, config_text, template_content, and vars_content as input (found in SKILL.md).
Boundary markers: None explicitly defined in the tool instructions.
Capability inventory: Remote CLI command execution and configuration modification via NETCONF.
Sanitization: The server implements regex-based blocklists (block.cmd and block.cfg) to prevent destructive operations (e.g., system reboot/zeroize) and unauthorized credential modifications.

junos-network